[OLSR-users] Host vs. Network route propagation with OLSR

[Andreas Tønnesen]

Klosch,

Your understanding is correct.
This is based on the idea that end users does not need to be able
to access intermittent backbone nodes.

But it was just me airing some ideas that I had pretty much forgotten
all about - so no need to worry that this will behappening in the near
future ;-)

- Andreas


kloschi wrote:
> On Thu, 2006-07-20 at 22:42 +0200, Andreas Tønnesen wrote:
>   
>>> I'm not sure I understand all of your second paragraph. It seems to
>>> me you're suggesting that each link between two WRT nodes have it's
>>> own small subnet shared between the two nodes.  If so, you might have
>>>       
>> No - what I'm suggesting is that we still do multihop ad-hoc routing, but
>> host-routes are not added for the routers since none of the routers(or
>> clients) are interested in talking directly to other routers. Therefore
>> only HNA enries has to be filled into the routing tables, this will make
>> sure multihop routing sill works dynamically and is basically just olsrd
>> without host routes.
>>     
>
> Just for my understanding - would that mean the WRTs are not directly
> accessable by IP anymore? if so, imho that wouldn't be the best to do,
> because the nodes are getting more and more used to host small stuff,
> like wikis, nanoblogs etc. ..
>
> or am I just totally confused?? hehe, 35 degrees in shadow doesn't make
> thinking easier :)
>
> kind regards,
> kloschi
>